Movement generation by Humans and Robots: a dynamical …€¦ · Robotics, companies in New...

Movement generation by Humans and Robots: a

dynamical systems perspective: Introduction

Gregor SchönerInstitute for Neural Computation

Theory of Cognitive [email protected]

mailto:[email protected]

Human movement generation

homo habilis or homo faber ...

we are the skillful species..

fluent sequences of movement, linked on-line to sensory information

flexibility: multiple motor skills which can be adapted and be performed concurrently

excellent fast scene perception

fine manipulation skills

Human movement generation

posture/balance

locomotion

navigating: moving through space

stepping

rhythmic (dance, music)

whole body skills, sports

reach, grasp, manipulate

speech articulatory movement

involuntary... automatic

voluntary

object oriented

What is motor control?

... the neural processes underlying the movement of organisms...

not just any movement

bacteria, plants: tropisms

falling from a tower...

What is entailed in generating an object-oriented movement?

scene and object perception

movement preparation

movement initiation and termination

movement timing and coordination

motor control

degree of freedom problem

movementpreparation

timing

control


=> spans perception, cognition and control movement

preparation

timing

control


it is difficult to isolate any individual process

=> this is why movement is so hard to study

=> this is why it is critical to understand integration

movementpreparation

timing

control

in this course I will…take you through the component process

illustrate theoretical concepts relevant to understanding motor behavior

summarize neurophysiology

use robotics to illustrate ideas

[Martin, Scholz, Schöner. Neural Computation 21, 1371–1414 (2009]

movement timing

neuronal dynamics of virtual joint trajectory

muscle-joint model

biomechanics

movement preparation

back- coupling

decoupling

basal ganglia cerebellum

premotor cortex

spatial representation

scene representation

visual system

parietal cortex

motor cortex

proprioception

spinal cord

sensori-motor periphery

movement initiation

Neurophysiology of movement

three systems

the levels of movement generation

cortex

brain stem

spinal cord

2 modulatory systems

loop through basal ganglia and thalamus

loop through cerebellum and thalamus

[Kandel, Schartz, Jessell, Fig. 35-3,all figures are from the 3rd edition]

the motor cortex

provides direct input to muscles through cortico-spinal projects

and inputs to spinal circuits and to brain stem

the cortico-spinal projection

leads to effector activation when motor cortex is excitated

motor homunculus

[Kandel, Schartz, Jessell, Fig. 40-1]

the cortico-spinal projection

stronger longer-lasting stimulation may lead to activation of complete movements

(recruiting multiple areas into the loop)

[Graziano, 2006]

in the motor cortical areas

movement parameters are encoded

e.g. direction of end-effector movement

e.g., direction of force-vector

in the sense of broad tuning

motor cortex: movement direction

[Georgopoulos, A et al. ]

population codeeach neuron contributes its perferred direction as a vector, with length=its current firing rate

vector sum=population vector predicts the movement direction

premotor cortices

are involved in movement preparation

[Graziano, 2006]

[Cisek, Kalaska, 2005]

spinal circuits: central pattern generators

spinal networks enable activation, deactivation, switching and self-stabilized oscillation

with coordinated alternation between agonist and antagonist activation


CPGs generate rhythmic

locomotory motor patterns

descending signal from brain stem activates a intrinsic spinal pattern in cat


the brain stem

the brain stem regulates/modulates spinal cord motor circuits

in the control of posture, the brain stem integrates visual and vestibular information with somatosensory inputs

brain stem nuclei control eye and head movements

... “old, basal” functions

spinal cord: reflex loops

alpha-gamma reflex loop generates the stretch reflex



the stretch reflex acts as a negative feedback loop

37-12



as a result, muscles are “tunable springs”

\

force applied

agonist

antagonist

spinal cord: coordination

Ia inhibitory interneuron mediates reciprocal innervation in stretch reflex, leading to automatic relaxation of antagonist on activation of agonist


spinal cord: synergies

Renshaw cells produce recurrent inhibition, regulating total activation in local pool of muscles (synergy)


spinal cord

=> the periphery of the motor system contributes to movement coordination, timing, and control….

Google search

Robots

12 Apr 2015, 12:25 roboter - Google Search

Page 1 of 12https://www.google.com/search?q=roboter&client=safari&rls…i=vVUqVcypGaev7AbN0ICgAQ&ved=0CAoQ_AUoBA&biw=1099&bih=659

Web News Shopping Videos Images Search toolsMore SafeSearch

roboter +Gregor

=> Humanoids (or anthropomorphic) robots

12 Apr 2015, 12:25 roboter - Google Search

Page 2 of 12https://www.google.com/search?q=roboter&client=safari&rls…i=vVUqVcypGaev7AbN0ICgAQ&ved=0CAoQ_AUoBA&biw=1099&bih=659

fundamentally, all factory automatization is a form or robotics today: “programmable” machines…

industrial robots are actually more common today

examples of robots

other than humanoid or industrial

5. Industrial, Personal, and Service Robots 58

environment to a remotely located human operator. Robots are able to provide logistics support in office and

industrial environments by transporting materials (packages, medicines, or supplies) or by leading visitors

through hallways. Remotely controlled and monitored robots are also able to enter hazardous or unpleasant

environments. Examples include underwater remotely operated vehicles, pipe cleaning and inspection robots,

and bomb disposal robots. Some examples are shown in Fig. 5.5.

Figure 5.5. Examples of service robots.

The challenges in service and personal robotics include all the challenges for industrial robotics. Dexterous

manipulation and integration of force and vision sensing in support of manipulation is critical to the growth

of this industry. In addition, mobility is a key challenge for service robotics. The current generation of robots

is only able to operate on two-dimensional, even, indoor environments. Because service robots must be

mobile, there are challenges for designing robots that are capable of carrying their own power source.

Further, operation in domestic environments imposes constraints on packaging. Finally, service robots,

especially personal robots, will operate close to human users. Safety is extremely important. And because

interaction with human users is very important in service robotics, it is clear the industry needs to overcome

significant challenges in human-robot interfaces.

INTERNATIONAL ASSESSMENT

U.S.

Most of the industrial robotics industry is based in Japan and Europe. This is despite the fact that the first

industrial robots were manufactured in the U.S. At one time, General Motors, Cincinnati Milacron,

Westinghouse and General Electric made robots. Now, only Adept, a San Jose-based company, makes

industrial robots in the U.S.

However, there are a number of small companies developing service robots in the U.S. iRobot and Mobile

Robotics, companies in New England, are pioneering new technologies.

Europe

The two big manufacturers of industrial robots in Europe are ABB and Kuka. Over 50% of ABB is focused

on automation products and industrial robots are a big part of their manufacturing automation with annual

revenue of $1.5B. ABB spends 5% of their revenues on R&D, with research centers all over the world. As in

5. Industrial, Personal, and Service Robots 58

environment to a remotely located human operator. Robots are able to provide logistics support in office and

industrial environments by transporting materials (packages, medicines, or supplies) or by leading visitors

through hallways. Remotely controlled and monitored robots are also able to enter hazardous or unpleasant

environments. Examples include underwater remotely operated vehicles, pipe cleaning and inspection robots,

and bomb disposal robots. Some examples are shown in Fig. 5.5.

Figure 5.5. Examples of service robots.

The challenges in service and personal robotics include all the challenges for industrial robotics. Dexterous

manipulation and integration of force and vision sensing in support of manipulation is critical to the growth

of this industry. In addition, mobility is a key challenge for service robotics. The current generation of robots

is only able to operate on two-dimensional, even, indoor environments. Because service robots must be

mobile, there are challenges for designing robots that are capable of carrying their own power source.

Further, operation in domestic environments imposes constraints on packaging. Finally, service robots,

especially personal robots, will operate close to human users. Safety is extremely important. And because

interaction with human users is very important in service robotics, it is clear the industry needs to overcome

significant challenges in human-robot interfaces.

INTERNATIONAL ASSESSMENT

U.S.

Most of the industrial robotics industry is based in Japan and Europe. This is despite the fact that the first

industrial robots were manufactured in the U.S. At one time, General Motors, Cincinnati Milacron,

Westinghouse and General Electric made robots. Now, only Adept, a San Jose-based company, makes

industrial robots in the U.S.

However, there are a number of small companies developing service robots in the U.S. iRobot and Mobile

Robotics, companies in New England, are pioneering new technologies.

Europe

The two big manufacturers of industrial robots in Europe are ABB and Kuka. Over 50% of ABB is focused

on automation products and industrial robots are a big part of their manufacturing automation with annual

revenue of $1.5B. ABB spends 5% of their revenues on R&D, with research centers all over the world. As in

[photo credits: WTEC final report 2006]

simple, single-task autonomous vehicles

some of our own autonomous

vehicles

outdoor vehicles

2. Robotic Vehicles 8

Figure 2.1. NASA Mars Rover (NASA Jet Propulsion Laboratory (JPL)).

Another example of a hostile and hazardous environment where robotic vehicles are essential tools of work and exploration is the undersea world. Human divers may dive to a hundred meters or more, but pressure, light, currents and other factors limit such human exploration of the vast volume of the earthLs oceans. Oceanographers have developed a wide variety of sophisticated technologies for sensing, mapping, and monitoring the oceans at many scales, from small biological organisms to major ocean circulation currents. Robotic vehicles, both autonomous and ROV types, are an increasingly important part of this repertoire, and provide information that is unavailable in other ways. Figure 2.2 shows an autonomous underwater vehicle (AUV) called ASTER under development at Institut Français de Recherche pour l'Exploitation de la Mer (IFREMER), the French National Institute for Marine Science and Technology. ASTER will be used for coastal surveys of up to 3,000 meters in depth and is capable of carrying a wide variety of instrumentation for physical, chemical, and biological sensing and monitoring. In United States research, the evolution of remotely operated vehicles for deep ocean exploration enabled the discovery of the sunken Titanic and the ability to explore that notable shipwreck.

Figure 2.2. IFREMER ASTER autonomous underwater vehicle.

In addition to space and oceans, there are many applications where human presence is hazardous. Nuclear and biological contamination sites must often be explored and mapped to determine the types and extent of contamination, and provide the basis for remediation. Military operations incorporate many different autonomous and remotely operated technologies for air, sea, and ground vehicles. Increasingly, security and defense systems may use networks of advanced mobile sensors that observe and detect potential events that may pose threats to populations.

In a second class of applications, robotic vehicles are used in routine tasks that occur over spaces and environments where machine mobility can effectively replace direct human presence. For example, large-

Arthur Sanderson, George Bekey, Brian Wilcox 9

scale agriculture requires machines to cultivate, seed, irrigate, and harvest very large areas of terrain. The ability to track an autonomous vehicle using global positioning systems (GPS), sensing the soil and plant conditions in the field, encourages the implementation of robotic vehicles for agricultural or EfieldF applications. Figure 2.3a shows an example of an agricultural robotic vehicle under development in the United States. Figure 2.3b shows a large autonomous mining haul truck developed in Australia.

(a) (b)

Figure 2.3. Agricultural robotic vehicle (Int Harv, U.S.) (a). Mining haul truck (ACFR, Australia) (b).

Similar challenges occur in areas of environmental monitoring, where mobile vehicles may move through air, water, or ground to observe the presence of contaminants and track the patterns and sources of such pollutants. In large manufacturing facilities, mobility is essential to transport components and subassemblies during the manufacturing process and a variety of robotic guided vehicles are utilized in these domains.

Figure 2.4. IBOT advanced wheel chair (DEKA, U.S.).

A third class of applications of robotic vehicles occurs in the support of personal assistance, rehabilitation, and entertainment for humans. A robotic wheelchair may provide mobility for a human who would otherwise not be able to move about. The integration of sensors, computational intelligence, and improved power systems have made such personal robotic aides increasingly capable and practical for everyday use. An example of a wheelchair that utilizes emerging robotic technologies for guidance and balance is shown in Figure 2.4. More details on medical robotics and robotic aids to the handicapped will be described in Chapter 6.

Other examples of such personal aides include vehicles that support elderly care through feeding, household tasks, and emergency notification. Many daily household tasks may benefit from enhanced mobile robotics, and there are rapid commercial developments of vacuum cleaners and lawn mowers that utilize advanced sensor and navigation systems. Also, advanced entertainment systems will incorporate robotic vehicles including locomotion of humanoids and biomimetic pets that entertain and provide interactive companions. The Japanese development of humanoids and robotic pets with sophisticated locomotion systems, as shown in Figure 2.5, is a major topic of this international comparative study. More detailed examples of personal and entertainment robotic vehicles will be described in Chapter 5, and of humanoid robots in Chapter 4.

cars: autonomous driving

the University of Central Florida in Orlando.They’ve outfitted Knight Rider, a 1996 Sub-aru Outback that belonged to Harper’s wifeand has 99,257 miles (159,705 km) on it, withjust enough gizmos to get around thecourse—they hope. Instead of the spinning3D lidar, they use two lidars that see in onedirection and rock them back and forth. “Ifjust one wire falls off, something essential isnot going to work,” Harper says. Still, theteam made the final having invested only$130,000 in the project.

Robots, start your engines!Race day usually brings the intoxicatingsmell of high-octane fuel and the electri-fying scream of engines. But not here. At8:00 a.m., the robots leave the startingarea, one by one, like rental cars leaving alot. There’s a glitch. Interference from ajumbo TV screen knocks out the GPSreceiver of first qualifier, Boss, CarnegieMellon’s Chevy Tahoe. The team replacesthe unit and has to wait 30 minutes toregain the signal. Meanwhile, Odin, a FordEscape from Virginia Polytechnic Instituteand State Universi ty in Blacksburg;Junior, Stanford’s Volkswagen Passat; andthe others head out, hesitating and swerv-ing as if driven by octogenarians. After ahalf-hour, all 11 robots—plus their chasecars and 37 other cars—are on the road.

There’s only one curve from which toglimpse the robots, so DARPA has hired ahelicopter and is televising the event on threehuge screens in a vast tent. Jamie Hyneman andGrant Imahara of the geeky cable-televisionreality show Mythbusters provide commen-tary. It’s like watching a hybrid of a NASCAR

race and the infamous O. J. Simpson low-speed police chase.

Each robot has to complete three “mis-sions” comprising six or seven “submis-sions,” such as parking in exactly the rightspace in a lot, traversing an off-road passage,or navigating between two places. After eachmission, the robots return to the start area todownload the specifications for the next, andeach machine must travel 60 miles (97 kilo-meters) in less than 6 hours.

At first, the action comes fast and heavy.An hour into the race, TerraMax, the hulkingvehicle entered by military contractorOshkosh Truck Corp. in Wisconsin, turnstoward a pillar and gets stuck staring at it.Forty-five minutes later, Central Florida runsstraight toward a house. Caroline, the robotfrom Team CarOLO, the other German squad,collides with MIT’s Talos and loses sensors.By 11:00 a.m., five robots have either failedor been disqualified.

Then things settle down. The remainingrobots’ “personalities” emerge. CarnegieMellon’s Boss zooms confidently away fromstops, a hard charger like team leader Whit-taker. Stanford’s Junior glides aroundsmoothly, so much so you hardly notice it.MIT’s Talos is aggressive in traffic—it alsoclips Cornell’s Chevy Tahoe, Skynet—butskittish off-road, stopping and starting like acat creeping down a steep slope.

Around 1:30 p.m., three teams have nearlycompleted their missions, and spectatorsswarm back to the grandstands. At 1:42, Stan-ford cruises across the finish line, followed aminute and a half later by Carnegie Mellon.Upstart Virginia Tech cruises home third—even without the 3D lidar. “We knew we were

good,” says Virginia Tech’s AlfredWicks. “We’d done our home-work.” The University of Pennsyl-vania’s Toyota Prius, Little Ben,straggles in an hour later. Some-time past 3:30 p.m., MIT slips injust before Cornell.

The outcome seems obvious.Carnegie Mellon spotted Stan-ford and Virginia Tech a 20-minute head start and made upalmost all of it. It seems the vic-tory should be theirs. DARPAofficials will make the final call,however. And, some participantsgrumble, DARPA never fullyexplains its judgments.

Make it out to …But the next morning bringsno surprises. Carnegie Mellonwalks off with the win. Stanford

takes second and $1,000,000, Virginia Techtakes third and $500,000. “There’s tremen-dous satisfaction in what the whole f ieldaccomplished,” Whittaker says. “That was aday that stunned the world.” DARPA Direc-tor Anthony Tether also gushes. “Quitefrankly, I watched these things and I forgotafter a while that there wasn’t anybody inthere,” he says. “It’s a historic day—’bot on’bot for the first time!”

Maybe there’s something to the grandioserhetoric. Now only a Luddite could doubt thatsoon cars will guide themselves, at least in apinch to avoid collisions. In fact, the technol-ogy already seems ripe for low-risk applica-tions, such as automating farm equipment,and the leading teams are pushing to com-mercialize their software. “I think it’s goingto come in bits and pieces,” says CharlesReinholtz, leader of the Virginia Tech teamand an engineer at Embry-Riddle Aeronauti-cal University in Daytona Beach, Florida.

Ironically, the success of the Urban Chal-lenge could reduce the chances that DARPAwill stage another competition. “DARPAnever finishes anything,” Tether says. “All wedo is show that it can be done” in the hope thatindustry takes over and pushes further devel-opment. Clearly, when it comes to makingrobotic cars, the Urban Challenge has shownthat it is possible.

Still, many engineers are eager for anothercompetition. Their robots aren’t nearly readyfor the open road, they say, and many alreadyknow what they would like to see in the nextchallenge: a contest for autonomous cars thatmust communicate and work together. Sud-denly, that doesn’t seem quite so absurd.

–ADRIAN CHO

www.sciencemag.org SCIENCE VOL 318 16 NOVEMBER 2007 1061

CR

ED

IT: D

AR

PA

NEWSFOCUS

After you. Stanford’sJunior and Virginia

Tech’s Odin negotiatean intersection.

Published by AAAS

on

Nov

embe

r 24,

200

7 w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

legged robotsC. Site Reports-Europe 230

Figure C.58. The walking machines built by DillmannCs group.

Up to now 10 six-legged walking machines have been built and distributed: three to museums and seven to universities or research groups in Germany and Europe.

Medical and Surgical Robotics

Several projects involve cooperation with medical personnel on visual identification of target areas for surgery on the skull, estimation of spine and neck muscle properties to determine the extent of whiplash injuries, and automatic calibration of medical instruments with imagery for image-guided surgery.

COLLABORATIVE PROJECTS

Professor Dillmann has an extensive network of collaborators on many projects. The big projects are:

• Collaborative Research Center on Learning and Cooperating Multimodal Humanoid Robots SSFB588,U which was established by the DFG in 2001 and will be run for 12 years (coordinator: Rüdiger Dillmann)

• Cogniron Project (funded by the EU, led by Dr. Raja Chatilla at the Laboratoire dCAnalyse et dCArchitecture des Systèmes (LAAS), Toulouse)

• PACO-PLUS (funded by the EU, coordinator: Rüdiger Dillmann)

Other collaborative research projects in Germany include:

• Robots in Surgery (SFB 414 in Karlsruhe) • Telepresence and Teleaction Systems (SFB 453 in Munich) • Autonomous Dynamic Walking (Munich) • Situated Artificial Communicators (SFB 360 in Bielefeld) • Robotic Systems for Handling and Assembly-High Dynamic Parallel Structures with Adaptronic

Components (SFB 562 in Braunschweig) • Spatial Cognition: Reasoning, Action, Interaction (Transregio SFB 6023 in Bremen/Freiburg) • Cognitive Cars (SFB in Karlsruhe/Munich)

snakes, crawlers, climbers

C. Site Reports-Europe

229

Figure C.56. Mobile robot platforms in Dillman>s laboratory. Two SwissLog products are shown on the

extreme right.

Pipe Inspection Robots

Prof. Dillmann>s lab has developed several articulated, snake-type robots for pipeline inspection. Some are now commercially available and used to inspect water pipes and oil pipelines (including the Alaska pipeline).

Figure C.57. Inspection robot.

Current work is concentrated on enabling the system to work in an unstructured environment. A multi-articulated system with six links will be used for inspection tasks in sewer pipelines.

Legged Locomotion

There are labs dedicated to the development of control systems for four- and six-legged robots, as well as bipeds. The emphasis appears to be in the application of artificial muscles (McKibben-type muscles with a rubber shield), reduction of size and weight, and joint design. Historically, they have fabricated several of the Lauron-type six-legged machines usually associated with Friedrich Pfeiffer at the Technical University of Munich (TUM). Apparently there has been a long-term cooperative effort between the two labs.

7. Networked Robots 74

exploiting the efficiency that is inherent in parallelism. They can also perform independent tasks that need to be coordinated (for example, fixturing and welding) in the manufacturing industry.

Networked robots also result in improved efficiency. Tasks like searching or mapping, in principle, are performed faster with an increase in the number of robots. A speed-up in manufacturing operations can be achieved by deploying multiple robots performing operations in parallel, but in a coordinated fashion.

Perhaps the biggest advantage to using the network to connect robots is the ability to connect and harness physically-removed assets. Mobile robots can react to information sensed by other mobile robots in the next room. Industrial robots can adapt their end-effectors to new parts being manufactured up-stream in the assembly line. Human users can use machines that are remotely located via the network. (See Fig. 7.3.)

The ability to network robots also enables fault-tolerance in design. If robots can in fact dynamically reconfigure themselves using the network, they are more tolerant to robot failures. This is seen in the Internet where multiple gateways, routers, and computers provide for a fault-tolerant system (although the Internet is not robust in other ways). Similarly, robots that can KplugL and KplayL can be swapped in and out, automatically, to provide for a robust operating environment.

Figure 7.2. Robotic modules can be reconfigured to KmorphL into different locomotion systems including a

wheel-like rolling system (left), a snake-like undulatory locomotion system (right), a four-legged walking system (bottom).

Finally, networked robots have the potential to provide great synergy by bringing together components with complementary benefits and making the whole greater than the sum of the parts.

Applications for networked robots abound. The U.S. military routinely deploys unmanned vehicles that are reprogrammed remotely based on intelligence gathered by other unmanned vehicles, sometimes automatically. The deployment of satellites in space, often by astronauts in a shuttle with the shuttle robot arm, requires the coordination of complex instrumentation onboard the space shuttle, human operators on a ground station, the shuttle arm, and a human user on the shuttle. Home appliances now contain sensors and

underwater vehicles, ships


Figure 2.1. NASA Mars Rover (NASA Jet Propulsion Laboratory (JPL)).

Another example of a hostile and hazardous environment where robotic vehicles are essential tools of work and exploration is the undersea world. Human divers may dive to a hundred meters or more, but pressure, light, currents and other factors limit such human exploration of the vast volume of the earthLs oceans. Oceanographers have developed a wide variety of sophisticated technologies for sensing, mapping, and monitoring the oceans at many scales, from small biological organisms to major ocean circulation currents. Robotic vehicles, both autonomous and ROV types, are an increasingly important part of this repertoire, and provide information that is unavailable in other ways. Figure 2.2 shows an autonomous underwater vehicle (AUV) called ASTER under development at Institut Français de Recherche pour l'Exploitation de la Mer (IFREMER), the French National Institute for Marine Science and Technology. ASTER will be used for coastal surveys of up to 3,000 meters in depth and is capable of carrying a wide variety of instrumentation for physical, chemical, and biological sensing and monitoring. In United States research, the evolution of remotely operated vehicles for deep ocean exploration enabled the discovery of the sunken Titanic and the ability to explore that notable shipwreck.

Figure 2.2. IFREMER ASTER autonomous underwater vehicle.

In addition to space and oceans, there are many applications where human presence is hazardous. Nuclear and biological contamination sites must often be explored and mapped to determine the types and extent of contamination, and provide the basis for remediation. Military operations incorporate many different autonomous and remotely operated technologies for air, sea, and ground vehicles. Increasingly, security and defense systems may use networks of advanced mobile sensors that observe and detect potential events that may pose threats to populations.

In a second class of applications, robotic vehicles are used in routine tasks that occur over spaces and environments where machine mobility can effectively replace direct human presence. For example, large-


Field Robotics

Robotic vehicles developed for both military and space applications are intended for use in rough terrain, that is, without roads or cleared areas. In this context, the experience of off-road robotic vehicles in the U.S. has also provided a basis for research in field robotics, the application of robotic vehicles to other unstructured domains, such as agriculture, mining, construction, and hazardous environments. In addition, U.S. industrial companies active in these areas have invested in prototype developments for these applications. Figure 2.3 is an example of these prototype vehicles.

Undersea Robotics

The United States has supported research in several different types of applications of underwater vehicles. These include:

a. Military and Defense Applications As described in KMilitary and Defense Systems,L U.S. defense technologies have included many fundamental prototypes and products that provide both ROV and AUV technology for the military. Figure 2.9 shows several of these vehicles.

b. Coastal Security and Environmental Monitoring Systems

AUV systems may be used as surveillance and observance of systems with both defense and environmental implications. Figure 2.10 shows an overview of the Autonomous Oceanographic Sensor Network (AOSN) systems, deployed as an experiment at the Monterey Bay Aquarium Research Institute (MBARI) in California, which integrate many different robotic and sensor resources.

Figure 2.10. Advanced Oceanographic Sensor Network

(MBARI, U.S.).

Figure 2.11. HROV (Hybrid ROV) project (Johns Hopkins University

(JHU) and Woods Hole (WHOL), U.S.).

c. Scientific Mission and Deep Ocean Science

AUV and ROV technologies are the only means to actively explore large portions of the ocean volume. The study of ocean currents, ocean volcanoes, tsunami detection, deepsea biological phenomena, and migration and changes in major ecosystems are all examples of topics that are studied with these systems. Several of the major scientific laboratories in the world are located in the U.S. and are leaders in these fields. A new project, HROV, is funded by the National Science Foundation (NSF) to develop a new hybrid remotely operated vehicle for underwater exploration in extreme environments, capable of operation to 11,000 meters depth as shown in Figure 2.11.

airborne robotsArthur Sanderson, George Bekey, Brian Wilcox 11

Table 2.1 Types of robotic mobility systems

RESEARCH CHALLENGES

Historically, there have been examples of technologies that could be controlled through remote mechanical linkages (e.g. mechanically coupled manipulators for handling dangerous chemicals), and other technologies that provided pre-programmed motions (e.g. missiles and torpedoes). However, only with the development of microelectronics and embedded computation has it been possible to design systems that combine both mobility and autonomy. Four major research challenges have dominated these developments, and they continue to represent the key themes observed in this international study:

Mechanisms and Mobility

As described above, both engineering and biomimetic approaches have been taken to design mobile robotics vehicles, and current research efforts continue to follow both of these strategies. Key research themes include:

Principles of Motion

Basic studies of kinematics and dynamics of motion in all domains (ground, air, and water) continue to examine fundamental issues of devices that contact and interact with the forces around them. A primary example of this work is the study of bipedal locomotion and the distinction between Iquasi-staticK walking and IdynamicK walking. Algorithms used in recent full humanoid prototypes exhibit very sophisticated motion and balance, but still do not achieve all of the characteristics of human dynamic balance. New theories and experiments continue to impact this research. Similarly, such studies have a direct effect on different walking patterns, such as trotting and running gaits, and how these may be executed on two-legged and multi-legged robotic vehicles.

Materials Properties and Design

Materials considerations are also of primary interest for new mechanisms, and uses of light and strong materials, with controllable compliance, are current research topics.


INTERNATIONAL SURVEY

Robotic vehicles have been a principal theme of robotics research in many of the laboratories that were visited in this international survey. In many cases, the emphasis of types of vehicles, approaches to design, and the applications of interest have varied among these different international communities. This section summarizes these observations.

Research on Robotic Vehicles – United States

In the United States, research on robotic vehicles has emphasized work in the following five areas:

Military and Defense Systems

U.S. investment in robotic vehicle research has strongly emphasized the development of ground, air, and underwater vehicles with military applications. As shown in Figure 2.9, there have been significant accomplishments in these areas in which large development programs have resulted in capable and reliable vehicle systems. Many of these systems are deployed in a Lremotely-operatedN mode, that is, a human controller works interactively to move the vehicle and position based on visual feedback from video or other types of sensors. In addition, there is a strong emphasis on integration of autonomous probes and observers with other parts of the military tactical system. The integration of sophisticated computer and communications architectures is an essential feature of these systems, and the use of algorithms such as SLAM to interpret complex scenes is an important contribution to these systems. The U.S. is generally acknowledged as the world leader in military applications of robotic vehicle technologies.

Figure 2.9. Examples of military and defense robotic vehicles.

Space Robotic Vehicles

The field of space robotics was identified as a topic for separate focus in this study and the major results of that effort will be presented in Chapter 3. In the context of vehicle technologies, the recent Mars rover programs have uniquely demonstrated perhaps the most successful deployment of robotics vehicle technologies to date in any domain of applications. The rovers have landed and explored the surface of Mars and have carried out important scientific experiments and observations that have dramatically enhanced human understanding of that planet and its natural history. This U.S. NASA effort has been the only successful demonstration of interplanetary vehicle space technology and is clearly recognized as the world leader in this domain.

u Integration of electronics into thejoint, leading to a modular design:This allows the design ofrobots of increasing kinematiccomplexity based on inte-grated joints as in the case ofthe DLR humanoid Justin.Moreover, one obtains a self-contained system, which iswell suited for autonomous,mobile applications.

u Full-state measurement in thejoints: As will be outlined inthe ‘‘Compliance Controlfor Lightweight Arms’’ sec-tion, our robots use torquesensing in addition to posi-tion sensing to implementa compliant behavior anda smooth, vibration-freemotion. The full-state mea-surement in all joints is per-formed at 3-kHz cycleusing strain-gauge-basedtorque-sensors, motor posi-tion sensing based on mag-netoresistive encoders, andlink side position sensorsbased on potentiometers(used only as additionalsensors for safety considerations).

u Sensor redundancy for safety: Positions, forces, andtorques are redundantly measured.

These basic design ideas are used for the joints in the arms,hands, and torso of the upper body system Justin (Figure 1).Moreover, because the joints are self-contained, it is straight-forward to combine these modules to obtain different kine-matic configurations. For example, the fingers have been usedto build up a crawler prototype. Figure 2 shows the explodedviewof one LWR-III (DLR-LWR-III) joint.

Compliance Control for Lightweight ArmsIn the next two sections, the framework used to implementactive compliance control based on joint torque sensing issummarized. The lightweight design is obtained by using rel-atively high gear reduction ratios (typically 1:100 or 1:160),leading to joints that are hardly backdrivable and have alreadymoderate intrinsic compliance. Therefore, we model therobot as a flexible joint system. Thus, measuring the torqueafter the gears is essential for implementing high-perform-ance soft-robotic features. When implementing compliantcontrol laws, the torque signal is used both for reducing theeffects of joint friction and for damping the vibrations relatedto the joint compliance. Motor position feedback is used toimpose the desired compliant behavior. The control frame-work is constructed from a passivity control perspective bygiving a simple and intuitive physical interpretation in terms

of energy shaping to the feedback of the different state vectorcomponents.

u A physical interpretation of the joint torque feedbackloop is given as the shaping of the motor inertia.

u The feedback of the motor position can be regarded asshaping of the potential energy.

(a) (b) (c)

(d) (e) (f)

Figure 1. Overview of the DLR Robots. (a) The DLR-LWR-III equippedwith the DLR-Hand-II. (b) TheDLR-KUKA-LWR-III that is based on the DLR-LWR-III. (c) The DLR humanoidmanipulator Justin. (d)The DLR-Hand-II-b, a redesign of the DLR-Hand-II. (e) The DLR-HIT hand, a commercialized versionof the DLR-Hand-II. (f) The DLR-Crawler, a walking robot based on the fingers of the DLR-Hand-II.

Link Position SensorCross Roller Bearing

Power Converter UnitJoint- and MotorcontrollerBoard Power Supply

Carbon FiberRobot Link

DLR RoboDrive withSafety Brake andPosition Sensor

Harmonic DriveGear Unit

Torque Sensor withDigital Interface

Figure 2. The mechatronic joint design of the DLR-LWR-III,including actuation, electronics, and sensing.

IEEE Robotics & Automation MagazineSEPTEMBER 2008 21

Robert Ambrose, Yuan Zheng, Brian Wilcox 47

Having hands will be essential in the early advancement of this research, since the learning and association of knowledge with objects will be done in the robotCs own terms, with the way a tool feels when grasped stored in sensori-motor space. Key advances in dexterous hands include tactile skins, finger tip load sensing, tendon drive trains, miniature gearing, embedded avionics, and very recent work in low-pressure fluid power systems. The fundamental research in biologically inspired actuators will likely transform the nature of this domain in the next ten to fifteen years.

Figure 4.10. Dexterous arms at DLR, NASA and UMASS.

Hands must be well-sized and integrated with their arms for best effect. One of the challenges that has made entry into this research domain difficult is the small number of arm options available to the researcher, and the corresponding high cost of such systems. There are few small, human-scale arms able to be integrated for mobile applications, and most of these have low strength. Most humanoid arms are low quality, have <6 degree-of-freedom (DOF)-positioning systems with backlash and little power that appear almost as cosmetic appendages. The AIST HRP2 system is one of the few bipedal humanoids that has strong arms, and the limbs can be used to help the robot get up from a prone position.

Figure 4.11. Strong dexterous arms at AIST Tsukuba, NASA and DLR.

The best arms in the field have integrated torque sensing, and terminal force-torque sensors that allow for smooth and fine force control. The arms have 7+ DOF, and are able to handle payloads on the order of 5 kg or higher. They have embedded avionics allowing for dense packaging and modular application.

Mobile Manipulation

Mobile manipulation is achieved when combining a lower body able to position itself with ease, and a dexterous upper body able to perform value-added work. While this combination is not necessarily humanoid, people are ideal examples of mobile manipulators. Active balancing bases or legs have small footprints, allowing their upper limbs to get close to the environment, while maneuvering in tight urban environments. Dual and dexterous upper limbs offer primate-like workspace and grasping abilities that can work with the interfaces and objects in those same urban environments. This class of machine can redistribute force and position control duties from lower bodies to upper bodies, where differences in drive trains and sensors offer complimentary capabilities. Pioneering work in this discipline was done at Stanford,

robotic manipulators, hands

some of our own robotic manipulators

mobile robot manipulators C. Site Reports-Europe 198

the rotation, finding the axis of rotation and complying. This concluded a set of six real-time demonstrations, which is unusual. The fact that the six systems were of such a high caliber should be noted.

Figure C.28. Dexterous arm on mobile base, opening door (left), robot passing through doorway (right).

our own mobile robot manipulator

[Arnold: 1998-2000]

In this course

we will refer to

vehicles

robotic arms with a vision sensor

auto-nomos: giving laws to oneself

minimally: autonomous robots generate behavior based on sensory information obtained from their own on-board sensors

in contrast to industrial robots that are programmed in a fixed and detailed way

autonomous robotics

but: even an industrial robot uses autonomous control to reach its programmed goals…

=> autonomy is expected to go beyond control, include decisions=qualitative change of behavior

e.g. avoid obstacle to the left vs. to the right

e.g., reach for one object rather than another

autonomous robotics

but: we do not expect autonomous robots to just do whatever “they want”… we expect to give them “order”

autonomous robotics

autonomy as a “programming interface”:

give instructions to a robot at a high level, in regular human language and gesture in a shared environment…

… and let the autonomous robot deal with the “details” of how to achieve goals

autonomous robotics

why autonomous robots?

asked my 19 year old son…

“I don’t know, to clean up, to serve drinks … but they are just generally cool /…

… (after some hesitation)… in the military

why autonomous robots?

at home, in the work place

collaborate with human users

assistance robotics

toy/entertainment/animation

including therapy (autism)

the “ideal” application because desire to remove human agent from the scene is consensual …

much US research

military, fire fighting, rescue





















B. Site Reports-Asia 110

The company also introduced three other products: the Basic Robot Education System, the Research and

Education Robot, which included various versions as mentioned earlier (Figure B.10 shows one of them),

and Military Robot. For the latter, Hanool has developed an Anti-Terror Robot called Hanuri-T-ML as shown

in Figure B.11. Like the first three robots, the Research and Education robot and the Anti-Terror robot are all

mobile robots equipped with various sensors for navigation and operation according to the purpose which

they were designed for.

Figure B.10. Research and Education Robot. Figure B.11. Military Robot.

SUMMARY

Hanool Robotics is the only small company the assessment team visited in this Japan-Korea trip. From the

visit, we can see that Korea has taken the new robotics industry seriously. Not only has the government set

up a policy defining Intelligent Service Robots as one of the ten growth engines of the economy, but in

addition, many small companies have started to develop new robot products. Hanool represents one of the 50

service robot companies in Korea that have collectively developed 20 robot product lines, of which Hanool

has two. It appears that Hanool is one of the more successful ones among all the start-up companies.

From the five products, we saw that the intelligence technology has important uses and great potential in the

service robots. We saw a great deal of high technologies used in service robots. Although some of them were

seen by team members earlier in North America, the development and implementation of the technologies for

application purposes by Hanool Robotics and others is impressive and commendable. In this regard, Japan

and Korea, especially Korea, are ahead of many countries. Experts in Korea expect the robot market to grow

based on three reasons; half of the households in South Korea reside in apartment-type homes that have

structured environments; Korea has a widespread network infrastructure that can support the operation of

robots; and people have a positive attitude towards the robot market. The start-up robot companies expect the

market of service robots to grow like that of mobile telephones in the near future. The team had to agree after

the visit to Japan and Korea.

may a military robot decide autonomously to shoot

…. navy ships do that already…

may a autonomous car decide between avoiding a pedestrian and preventing danger for car occupants?

fundamental problem: off-loading decisions from user to designer …

(robot ethics…interesting topic)

autonomous robotics as a “playground” of research

modern engineering models systems, treating the remainder stochastically…. autonomous robotics act in natural environments that are difficult to model

autonomous robotics: highly interdisciplinary

modern engineering uses modular design,that limits the range over which modules interact/interfere…autonomous robotics: requires system integration

autonomous robotics as a “playground” of research

robotics vs. human movement

shared functions, constraints

standard approaches are very different

but we will look at neural principles that can be used to build autonomous robots

the motor control problem is very different:

servo-control with very high stiffness/precision vs. soft spring-like control in humans

but: interest in compliant robots

e.g., grasping, changing the demands on perception by using spring-like actuators

autonomous robots as demonstrations of neural function

neural process models: capable of generating the modeled behavior based on real or simulated sensory information …

proof of function as a source of heuristics

discover problems that are often overlooked

the problem of synthesis or integration…

discover non-problems that need not be solved to achieve a function

Movement generation by Humans and Robots: a dynamical …€¦ · Robotics, companies in New...

Documents

Transcript of Movement generation by Humans and Robots: a dynamical …€¦ · Robotics, companies in New...